AI is a highway, not a slot machine — stop trading the noise

Why this episode exists

The episode opens with Muzamil thanking the audience for the response to the recent long-form videos and explaining, in plain terms, why he is moving away from the tightly structured, shorter format. The structured pieces travel well on the algorithm, but, in his own words, “a lot of things get lost in translation.” He wants the room for “longer form, honest, raw, authentic conversations” — the kind where the argument can actually breathe.

He then names the wedge for the next stretch of work. He wants to go deep on AI. Not the headline AI of new model launches and job-loss panics, but the underlying current. He has spent the last year travelling aggressively across the United States, sitting in rooms, having the kind of conversations that do not show up in a feed. The conviction he keeps returning to is simple: Pakistanis, at home and abroad, can participate in this cycle and make serious money doing it. “It’s not a matter of if, it’s a matter of when. And I also think we are early this time around.”

The overseas Pakistani moment

Before he gets to bubbles, Muzamil pauses on a structural observation that frames the rest of the episode. Previous growth spurts — the ones India and the Philippines rode — were largely missed by Pakistanis abroad. The reason, he argues, was not capability. It was posture. The first wave of any immigrant community keeps its head down. Confidence is missing. People stay on H-1Bs, work for somebody else’s company, and rarely move into visionary roles.

What he is seeing now is different. There is a generation of Pakistanis who climbed the corporate ladder inside US tech, reached VP and beyond, took the stock options, and have moved up Maslow’s hierarchy to the point where they can think about legacy rather than security. “This time, for the first time, I have seen it,” he says. The visionaries will come from this private side. The everyman will build the businesses on top. And that, in his reading, is the substrate on which a Pakistani participation in the AI wave can actually be built.

The mismatch the internet is selling

Muzamil names the trap directly. The dominant narrative online holds two contradictory ideas at the same time: AI is a bubble, and AI is going to change everything. People are anxious, they are tense, they want to do something, and his comment sections, he notes, fill up with the same question — “sir, koi pathway bata dein, koi direction de dein.”

His read is that the contradiction is a feature, not a bug, of the attention economy. “I feel like a lot of people are getting distracted by noise on the internet about AI. They are being hypnotised, because our attention economy runs on opium — and the opium is fear.” Every new model, every new product launch, is repackaged as a threat. The result is a generation that is too anxious to focus and is trying everything at once instead of championing one thing.

The rest of the episode is an attempt to give people a frame for resolving the mismatch without abandoning either claim. To do that, he walks through three historical bubbles.

Railway mania — when the first highway was laid

The first stop is Britain’s railway mania. The point Muzamil keeps hammering is that the railway was not a fake technology. It was groundbreaking. It compressed time and distance, and it changed how a country could function. But the narrative around it ran ahead of the deployment, the stock market priced in fantasy, and roughly two hundred companies eventually went bankrupt in the consolidation that followed.

The detail he wants the listener to hold is what happened on the other side of that crash. For the next fifty years, the most valuable companies in the world were railway companies. America’s “gilded age” families — the Rockefellers, the Carnegies — built their fortunes on this technology. The bubble bursting did not invalidate the highway. It just cleared out the people pricing it wrong.

The 1920s, electricity, cars, and the Great Depression

The second case is the 1920s. The decade was, again, a genuine technological revolution. Electricity was becoming everyday. Cars were going mass-market. Industrialisation was running at full speed. And the narratives ran wild — Muzamil mentions the wireless-electricity and teleportation stories that were circulating at the time, the same shape of speculation we see today.

He then anchors the crash with a single number. “One percent of Americans owned roughly twenty-four percent of the entire wealth in America” before the Great Depression. He flags this as a pattern rather than a one-off: “every major bubble did the same thing.” Inequality stretched, a permanent upper class seemed to be forming, and then the system broke. Global GDP fell by around thirty percent. Half of global trade was disrupted. It took the world a long time to recover.

The reason he wants this in the listener’s head is not despair. It is to puncture a specific fantasy being sold today — that Elon Musk and a handful of AI companies will simply own everything forever. “Historically, the world has never run that way. Whenever inequality has gone too far, the system breaks.”

The dot-com bubble — the closest mirror

The third and closest analogue is the dot-com bubble of 2001. Muzamil walks through the two-phase shape carefully. The first phase was telecom and infrastructure — Cisco being the headline winner, riding the build-out of the wires everyone would eventually connect on. The second phase was the consumer internet, where valuations decoupled from any underlying economics and the crash wiped out a generation of companies.

The post-crash narrative is the one he wants the audience to remember. The story flipped overnight. The internet had been “the future” the day before the crash and “a scam, don’t touch it, get into petroleum” the day after. The people who internalised that flipped narrative missed what came next. The companies that were eventually built on top of that same internet are, in his words, “not only ruling the world right now — they are the most powerful people in the world right now.”

Reconciling the bubble and the revolution

This is the synthesis the whole episode has been building toward. “You have to reconcile this noise of AI being a bubble and AI completely changing the world by understanding that there is a timeline deficit in between.” The two claims are not contradictory. They sit at different points on the same curve.

What that means in practice, Muzamil argues, is that the people selling fear and the people selling FOMO are both wrong in the same way. They are collapsing a thirty-year arc into a six-month trade. The right response is to step back from the tools-and-services churn and look at the underlying technology. “Focus on the underlying technology and its impact. Don’t try to build some tool, sell it, raise from an investor, flip it, and get rich. Realise what impact AI is actually bringing.”

The question he wants people to sit with is structural. If AI is genuinely lowering the cost of intelligence, what changes in the output? What new businesses, what new services, what new categories become possible when intelligence is cheap and widely available?

Infrastructure money vs outcome money

He is careful to distinguish two layers. The infrastructure layer — today’s Nvidia, yesterday’s Cisco — does make a lot of money during the build-out. That is real. But it is also, historically, where the crash hits hardest. The companies that compound for decades are the ones on the impact layer: outcome-focused, real-world-value-focused businesses that use the intelligence highway to do something specific for someone specific.

Muzamil flags that he will unpack what “outcome” and “real-world value” actually mean in coming videos. For this episode, the instruction is to stop staring at the layer everyone else is staring at.

Don’t try every tool — pick one and champion it

Tucked inside the macro argument is a piece of personal advice for the people writing to him in the comments. The anxiety, he says, is producing the wrong behaviour. People are trying every new tool that drops, hoping one of them will be the lottery ticket. That is precisely the opposite of what compounds. “Instead of putting your head down and focusing on one thing, championing it, growing incrementally — they are just trying everything.”

The discipline he is pointing at is small and unglamorous. Pick a thing. Stay with it. Let the curve do its work.

The highway and the restaurants

By the end of the conversation, Muzamil lands on the metaphor he wants people to sleep on. AI is a highway. It is the same shape of thing electricity built, that the telephone built, that the internet built — a piece of underlying infrastructure on top of which a range of industries gets to exist.

The question, then, is not “should I become an LLM engineer.” Not everyone needs to be technical. The question is: what gets built alongside the highway? “On the motorway, the people who opened restaurants at Bhera made serious money.” Same logic here. What services, what businesses, what categories can you build that sit on the AI highway and serve real people?

He closes by asking the audience to actually do the homework — go and study the previous bubbles, look at where the value eventually generated, which companies survived, which jobs disappeared and which emerged. And then, before sleeping, to sit with the highway question. If intelligence becomes widely available across society, what would you build? The episode ends not on a prediction, but on a prompt.